Maximum-intensity-linking for segmentation of fluorescence-stained cells

Fluorescence microscopy is a key technology for the development of new agents in the area of pharmaceutic biotechnology and an important element of computer aided diagnosis and prognosis of tumor cells. A critical component of automatic image analysis methods for cell screening applications is cell segmentation. That is, in order to determine any speci c cell features, the individual cells have to be segmented from each other and the background. We propose an approach called 'maximum-intensity-linking' for the segmentation of uorescence stained cells that is related to classic watershed methods. It is simple, robust and general enough to be easily adapted to changing segmentation requirements.